POOOLOTTI AND SANTI
Morton, D. J., Hoppel, C., and Cooper, C. (1968), Biochem. J . 107,377. Razin, S. (1972), Biochim. Biophys. Acta 265,241. Roy, A. B. (1971), Biochimie53,1031. Schnaitman, C., Erwin, V. G., and Greenawalt, J. W. (1967), J. Cell Biol.32,719.
Shimazaki, J., Horaguchi, T., Ohki, Y . ,and Shida, K. (1971), Endocrinol. Jap. 18,179. Sperry, W. M., and Webb, M. (1950), J . Biol. Chem. 187,97. Vallejo, C. G . ,and Lagunas, R. (1970), Anal. Biochem. 36,207. Wilson, J. D., and Gloyna, R. E. (1970), Recent Progr. Horm. Res. 26,309.
Model Studies of the Thymidylate Synthetase Reaction. Nucleophilic Displacement of 5-~-Nitrophenoxymethyluracils t Alfonso L. Pogolotti, Jr.,l and Daniel V. Santi*
Nucleophilic displacement reactions of 5-pnitrophenoxymethyluracil and its N-alkylated derivatives have been examined to provide insight into the mechanism by which thymidylate synthetase catalyzes hydride transfer from 5,lOmethylenetetrahydrofolate to the methyl group of thymidylate. All reactions appear to proceed by formation of highly reactive intermediates having a n exocyclic methylene group at the 5 position of the heterocycle rather than direct displacement ( S N ~of ) the leaving group. The driving force for the expulsion of the leaving group and formation of such intermediates may be provided by the N-1 anion, where possible, or by attack of a nucleophile at the 6 position of the heterocycle ABSTRACT:
A
minimal mechanism for the thymidylate synthetase catalyzed reductive methylation of dUMP1 to dTMP must involve at least two steps. This was recognized some time ago when Friedkin and coworkers (Friedkin and Kornberg, 1957; Pastore and Friedkin, 1962) proposed that condensation of CH2-H,folate with dUMP results in a 5-thymidylyl-H4folate inteimediate which subsequently undergoes disproportionation ria a 1,3-hydride shift to give the products dTMP and 7,8-H2folate (Figure 1). In the first step, an electrophilic substitution reaction occurs in which the methylene carbon of 5,10-CH2-H,folate replaces the hydrogen at the 5 position of dUMP without a change in oxidation level. The second step of this mechanism can best be described as a nucleophilic substitution at the incipient methyl group of dTMP by hydride,
when the 1 position is alkylated. Direct support for the proposed mechanisms was obtained by evaluation of secondary deuterium isotope effects of reactants possessing deuterium at the 5-methylene carbon or the 6 position of the heterocycle. The mechanism involving nucleophilic attack at the 6 position of the heterocycle is analogous to that observed in model studies of other reactions catalyzed by this enzyme, and permits us to propose a unified mechanism for catalysis which is supported by all chemical and biochemical data at hand. Discussion is presented which argues against the existence of a thymidylyl-tetrahydrofolate intermediate in the reaction pathway leading to products.
originating from the 6 position of the cofactor, and resulting in concomitant production of 7,8-H2folate. Wherever possible, radioisotope tracer experiments have verified salient features of this mechanism (Pastore and Friedkin, 1962; Blakley et al., 1963), but increasing knowledge of the chemistry of the components of this reaction has made it apparent that the enzymic reaction is much more complicated than originally proposed. Previous reports from this laboratory (Santi and Brewer, 1968, 1973; Santi et ai., 1970) dealt with the development of model systems which would help to elucidate the mechanism of the condensation of dUMP and the formaldehyde donor. These studies led to the conclusion that the reaction was initiated by nucleophilic attack at the 6 position of d U M P and resulted in activation of the 5 position toward a species of formaldehyde. Subsequent studies using the quasi-substrate, FdUMP, demonstrated the formation of isolable covalent f From the Department of Biochemistry and Biophysics and Departenzyme--FdUMP complexes (Santi and McHenry, 1972) in ment of Pharmaceutical Chemistry, University of California, San which a nucleophile of the enzyme was attached to C-6 of the Francisco, California 94143. Receiced MUJ 31, 1973. This work was supported by U . S. Public Health Service Grant No. CA-14394 from analog and strongly supported the congruence of the model the National Cancer Institute. systems. $ National Institute of Health Predoctoral Fellow, 1968-1971. A question which remains unanswered is how the second, or Present address: Department of Chemistry, University of Arizona, oxidation-reduction, stage of the reaction occurs. In view of Tucson, Ariz. 85721. 1 Abbreviations used are: dUMP, 2’-deoxyuridylic acid; dTMP, the poor leaving group potential of amines, it is difficult to deoxythymidylic acid; 5,1O-CHz-Htfolate, 5,lO-methylenetetrahydrofolic envision why the N-methylene group of an intermediate such acid; 7,X-Hzfolate, 7,8-dihydrofolic acid; Hafolate, tetrahydrofolic acid; as 5-thymidylyl-Hifolate would be susceptible to nucleophilic F d U M P , 5-fluoro-2’-deoxyuridylic acid; H M U , 5-hydroxymethyluracil; attack by hydride; to our knowledge, precedent for direct I M e H M U , 3MeHMU, and Me2HMU, 1-methyl-, 3-methyl-, and 1,3nucleophilic displacement at the a carbon of a tertiary amine is dimethyl-5-hydroxymethyluracil, respectively; NPMU, 5-p-nitrophenoxymethyluracil; IMeNPMU, 3MeNPMU, and MezNPMU, 1lacking. In addition, we wondered whether the nucleophilic methyl-, 3-inethyl-, and 1,3-din~ethyl-5-p-nitrophenoxyinethyluracil, recatalyst required in the initial stage of the reaction might also be spectively; lAPr-3MeNPMU, 1-(3-aminopropyl)-3-methyl-5-p-nitroinvolved in the oxidation-reduction step. With these inquiries phenoxymethyluracil; lAPr-3MeHMU, 1-(3-aminopropyI)-3-methyl-5in mind, we sought to study related reactions in a simpler chemhydroxymethyluracil.
456
B I O C H E M I S T R YV,O L . 1 3 ,
NO.
3, 1 9 7 4
MODEL STUDIES OF THYMIDYLATE SYNTHETASE
ical system which would mimic that which is catalyzed by the enzyme. We have recently observed that esters and ethers of 5-hydroxymethyluracil are unusually reactive toward nucleophilic displacement (Santi and Pogolotti, 1971), the former undergoing 0-alkyl rather than 0-acyl bond cleavage. In addition, when treated with hydride reagents, such compounds rapidly gave rise to corresponding thymine derivatives. In this report, we describe detailed investigations of the mechanisms of nucleophilic displacement reactions for derivatives of 5-pnitrophenoxymethyluracils, and offer these as chemical counterparts of the second step of the thymidylate synthetase reaction. Although the leaving group in the model system greatly differs from H,folate in structure and reactivity, we feel that the salient features of the reaction reside in the uracil heterocycle, and are retained regardless of the nature of the leaving group. From these studies, we are able to offer a reasonable chemical mechanism for the enzymic reaction which is consistent with all biochemical data reported to date. Experimental Section General. All materials were reagent grade unless otherwise specified. Dimethylformamide was dried over Linde 4A Molecular Sieves. Potassium carbonate was dried at 150" prior to use. Melting points were determined on a Mel-Temp apparatus and are corrected. Infrared spectra (KBr) were recorded on a Perkin-Elmer Model 337 spectrophotometer. Deuterium analysis was performed on an AEI-MS902 mass spectrometer equipped with a Mosley 7101B linear streak recorder and a Varian HA-I00 nuclear magnetic resonance (nmr) spectrometer in conjunction with a Varian Model C-1024 time averaging computer. Dideuterioparaformaldehyde was purchased from Merck and Co. Methanol-l-d (99.773 and D 2 0 (99.8z) were purchased from Stohler Chemicals. Sodium deuteroxide solutions were prepared by dissolving metallic sodium in deuterium oxide. Thin-layer chromatography (tlc) was performed on silica gel GFZ5, (Merck) plates and paper chromatography (ascending) was performed on Whatman No. 3MM strips. Spots were visually detected under short-wave ultraviolet (uv) light. Ninhydrin spray was employed to detect primary amines and 10% ethanolic HC104 was used to detect the monomethoxytrityl blocking group (Schaller et al., 1963). Elemental analyses were performed by Galbraith Laboratories, Knoxville, Tenn. The synthesis of 1APr-3MeHMU (Scheme I) was performed by the following sequence. SCHEMEI 0
3MeHMU
0
0
,CH,R
IAPr-3MeNPMU la, R = OH b. R = C1 c. R = OC,H,NO,
1,1-F4 H ,
I,IO-CH.FAH.
1: Reaction catalyzed by thymidylate synthetase showing the thymidylyl-H4folateintermediate proposed by Friedkin ; R = 5-phospho-2 '-deoxyribosyl.
FIGURE
I -[3-(p-Methoxytritylamino)pro~y~-3-methyl-5-hydroxymethyluracil (la). To 50 ml of chilled (5") CHCl3 containing 5.6 ml (40 mmol) of triethylamine and 2.19 g (10 mmol) of 1-amino-3-bromopropane hydrobromide was added 3.40 g (1 1 mmol) of p-anisylchlorodiphenylmethane over a 30-min period with vigorous stirring. The chilled solution was stirred an additional 30 min and then washed with 3 X 10 ml portions of 10% K H C 0 3 followed by 3 x 15 ml washings with H 2 0 . The organic layer was dried (MgSO,) and evaporated in vacuo at room temperature to give 3.89 g (9773 of a clear yellow oil which showed only trace impurities on tlc using petroleum ether-EtOAc (1O:l). The oil was chromatographed on a silica gel (Woelm) dry column (3.8 X 40 cm) (Loev and Goodman, 1967) using the same solvent system. The product was eluted with 5 X 10 ml portions of CHC13, dried (MgSOn), filtered, and evaporated in vacuo to give 3.60 g (90%) of a colorless oil. Nmr and infrared (ir) spectra were consistent with the assigned structure. The oil gave one uv absorbing spot on tlc using petroleum ether-EtOAc (10 :1) which upon spray treatment with 10% ethanolic HC10, produced an orange color. The product was used immediately in the subsequent reaction. A suspension of 3.90 g (25 mmol) of 3-methyl-5-hydroxymethyluracil and 3.46 g (25 mmol) of K2C03 in 75 ml of dimethylformamide was magnetically stirred at ambient temperature for 2 hr protected from moisture. After 2 hr, 10.1 g (25 mmol) of the above monomethoxytrityl derivative was added and the suspension was stirred for 5 days at ambient temperature protected from moisture. The dimethylformamide was evaporated in vacuo at room temperature. The solid residue was dissolved in 1.50 m1 of CHC13and washed with 5 x 20 ml portions of water. The organic layer was dried (MgS04) and evaporated in vacuo to give a white residue. Crystallization from benzene-petroleum ether gave 6.0 g (50.4z) of product; mp 92-93.5". A portion was recrystallized in similar fashion to give the analytical sample: mp 93-93.5"; h,,,(H20) 271 nm (pH 7). Tlc using EtOAc-petroleum ether (3:l) showed one spot which gave a positive spray test for the monomethoxytrityl group. Anal. Calcd for C29H31N304.2/3C8H6: C, 73.72; H, 6.55 N, 7.82. Found: C, 73.77; H, 6.60; N, 7.88. I -[3-(p-Methoxytrit~lamino)propy~-3-methylJ-chloromethyluracil (Zb).To a magnetically stirred solution (-5") of 0.970 g (2.0 mmol) of l a and 4.0 ml of triethylamine (29 mmol) in 20 ml of CHCll was slowly added 0.150 ml(2.1 mmol) of S0Cl2. After 30 min at -5" the solution was evaporated in vacuo to give a clear glass which was dissolved in a minimal amount of CHC13 and applied to a silica gel column (2 X 31 cm). Elution with EtOAc-petroleum ether (1:l) gave 0.700 g ( 7 1 z ) of BIOCHEMISTRY, VOL.
13,
NO.
3, 1 9 7 4
457
POGOLOTTI AND
SANTI
(IAPr-3MeHMU). The same method described for the product pure by tlc in the form of a clear glass. The high preparation of the corresponding p-nitrophenyl ether was reactivity of the 5-chloromethyl group made analysis difficult used. Starting with 0.300 g (0.62 mmol) of N-blocked 1APrand structure assignment was based on the absence of O H 3MeHMU (la), 0.1 38 g (82 %) of a colorless oil was isolated as stretch in the ir spectrum, a positive p-nitrobenzylpyrjdine test for active halogen (Epstein er al., 1955; Baker et al., 1966), a the acetate salt which showed one spot on tlc using EtOAc-HCOOH-HzO (7 :2 :1) as the solvent system. Elution from a positive spray test for the monomethoxytrityl group, and quantitative conversion to starting material upon treatment phosphocellulose (0.94 mequiv,/g) column (3 x 27 cm) ocwith aqueous triethylamine: vmRX 3200 (NH), 1710-1640 curred at 0.15 M HOAc using the same conditions described for (C-C, C=O), 1600, 1500. 760, 830 cm-' (phenyl). The prod1APr-3MeNPMU. Unlike lAPr-3MeNPMU, the product was stable at room temperature or in strongly basic solutions (pH uct was immediately used in the next reaction. I-[3-(p-Methoxytritvlomino)propyl]-3-methyl-5-p-nitrophe- >12). The product was characterized by homogeneity on the chromatography system discussed earlier, a positive ninhydrin noxymethylurocil ( I C ) . To a well-stirred solution of 2.28 g (14.2 test, and a negative test for the monomethoxytrityl group: mmol) of anhydrous sodium p-nitrophenolate and 0.328 g (2 A,,,(H,O) 268 ( 6 8120) (pH 2), 269 nm ( e 8190) (pH 10). mmol) of KI in 15 ml of dimethylformamide was added 0.700 g Uracil-5,6-d: in 98.5 isotopic purity was prepared by (1.42 mmol) of the blocked chloromethyluracil (lb). After catalytic hydrogenolysis of 5-bromouracil-6-d by the method stirring for 18 hr at ambient temperature, the solution of Parkanyi and Sorm (1963) with the exception that Dz gas was evaporated in vacuo (
